There is a growing policy focus in Australian higher education on quantitative research performance assessment. However, most of the analysis has addressed aggregate performance at the institutional level, an approach inconsistent with recent policy emphasis on diversity among universities, and one that ignores performance variations across disciplines. We use cluster analysis to classify one of the ten broad fields of education, that is, management and commerce. Using averaged and available data for 2000-2004 on various research measures, partial rankings are provided. Factor analysis is utilised to generate full-multidimensional rankings within the resulting clusters. Our results show that low total research output and poor average (per capita) performance are closely correlated. We hypothesize that a minimum scale of efficiency exists as a result of the need to cover research overheads and to embrace the benefits of communities of practice in disciplinary research.